Deployment: Overview

DataStream offers flexible deployment options to match your specific infrastructure requirements. Whether you're running on local hardware, cloud virtual machines, or managed services, DataStream can be deployed in a way that optimizes performance and minimizes operational overhead.

• To deploy locally, see the On Local section of the chapter.

• To deploy on cloud, see the relevant sections for Azure in the same chapter. Here, we will deploy on premises as server—for Windows, Linux, macOS, and Copilot+ (ARM64).

• To deploy as clusters, see the As Cluster section in the same chapter.

Once you have decided on your deployment model and chosen the location that suits you, use the relevant sections to install the binaries and make the required settings.

Models

DataStream supports the following deployment models:

Deployment Type	Description	Use Case
Local	Run DataStream on your own hardware	Development, testing, or small-scale deployments
Azure VM	Deploy on Azure Virtual Machines	Production workloads with custom configurations
Azure App Service	Run as a managed service	Simplified operations with automatic scaling
Azure Functions	Event-driven serverless deployment	Cost-effective processing of intermittent workloads
Azure Arc Extension	Deploy to hybrid or multi-cloud environments	Consistent management across diverse infrastructure

Prerequisites

Before deploying DataStream, ensure your environment meets these requirements:

• Hardware Requirements:

  • CPU: 2+ cores (4+ recommended for production)
  • RAM: 4GB minimum (8GB+ recommended for production)
  • Storage: 20GB minimum for the application

• Software Requirements:

  • Operating System: Linux (Ubuntu 20.04+, RHEL 8+), Windows Server 2019+
  • .NET Runtime: 6.0 or later

• Network Requirements:

  • Inbound ports: Based on configured collectors (e.g. 514 for Syslog)
  • Outbound connectivity to your storage and analytics services

Choosing the Right Model

Consider these factors when selecting a deployment option:

1. Data Volume: Higher volumes may require VM-based deployments for maximum performance
2. Operational Overhead: Managed services reduce maintenance but may have configuration limitations
3. Cost Model: Serverless options can be cost-effective for variable workloads
4. Integration Requirements: Some deployment options offer better integration with existing systems
5. Security Requirements: Different deployment models have different security boundaries

Process

Regardless of the deployment model, the general process follows these steps:

1. Prepare Environment: Set up the target infrastructure and dependencies
2. Configure DataStream: Define collectors, pipelines, and output destinations
3. Deploy Application: Install DataStream using the appropriate method
4. Verify Operation: Check logs and metrics to ensure proper functioning
5. Monitor and Scale: Adjust resources based on actual usage patterns

Next Steps

Choose the deployment model that best fits your needs from the navigation menu. Each section provides detailed, model-specific instructions for setting up DataStream.

For questions or assistance with deployment, contact our support team at support@virtualmetric.com.